Skip to content

README.md

Latest commit

 

History

History
124 lines (88 loc) · 3.99 KB

README.md

File metadata and controls

124 lines (88 loc) · 3.99 KB

AI Impact Analysis

This script analyzes a codebase to measure and report the impact of AI-generated code contributions. It provides detailed insights about AI vs human contributions, helping teams understand the role of AI in their development process.

Features

  • Repository Analysis: Automatically detects and analyzes git repositories:

    • Uses current directory if it's a git repo

    • Searches parent directories for a git repo

    • Falls back to cloning a specified repository if needed

    # Basic repository setup
repo_path = os.getcwd()
repo_config = RepoConfig.from_repo_path(repo_path)
repo_operator = RepoOperator(repo_config=repo_config)
project = ProjectConfig.from_repo_operator(repo_operator=repo_operator, programming_language=ProgrammingLanguage.PYTHON)
codebase = Codebase(projects=[project])

  • Comprehensive Statistics:

    • Total number of commits and AI vs human contribution percentages
    • Files with significant AI contribution (>50%)
    • AI-touched symbols and their impact
    • Detailed contributor breakdown (human and AI contributors)
    # Run the analysis
ai_authors = ["github-actions[bot]", "dependabot[bot]"]
results = analyze_ai_impact(codebase, ai_authors)

# Access statistics
stats = results["stats"]
print(f"Total commits: {stats['total_commits']}")
print(f"AI commits: {stats['ai_commits']} ({stats['ai_percentage']:.1f}%)")
print(f"Files with >50% AI: {stats['ai_file_count']} of {stats['total_file_count']}")

# View contributors
for author, count in results["contributors"]:
    is_ai = any(ai_name in author for ai_name in ai_authors)
    print(f"{'🤖' if is_ai else '👤'} {author}: {count} commits")

  • High-Impact Code Detection:

    • Identifies AI-written code that is heavily used by other parts of the codebase
    • Shows dependency relationships for AI-contributed code
    # Access high-impact AI symbols
for symbol in results["high_impact_symbols"]:
    print(f"Symbol: {symbol['name']} ({symbol['filepath']})")
    print(f"Used by {symbol['usage_count']} other symbols")
    print(f"Last edited by: {symbol['last_editor']}")

# View top AI-contributed files
for file_path, percentage in stats["top_ai_files"]:
    print(f"{file_path}: {percentage:.1f}% AI contribution")

  • Detailed Attribution:

    • Maps symbols to git history
    • Tracks last editor and complete editor history for each symbol
    • Flags AI-authored symbols
    # Get attribution information for a specific symbol
symbol = codebase.get_symbol("path/to/file.py:MyClass.my_method")

# Access attribution data
print(f"Last editor: {symbol.last_editor}")
print(f"Editor history: {symbol.editor_history}")
print(f"AI authored: {symbol.is_ai_authored}")

# Find all AI-authored symbols
ai_symbols = [s for s in codebase.get_symbols() if s.is_ai_authored]
for symbol in ai_symbols:
    print(f"AI symbol: {symbol.name}")

Output

The script generates:

  1. Console output with summary statistics
  2. Detailed analysis in ai_impact_analysis.json
  3. Attribution information added to codebase symbols

Usage

python run.py

The script will automatically:

  1. Initialize and analyze the codebase
  2. Process git history
  3. Generate attribution information
  4. Output detailed statistics

You can also visualize the AI impact analysis results using a dashboard. For setup and usage instructions, please see the documentation in the /dashboard subdirectory.

Symbol Attribution

After running the analysis, symbols in the codebase will have the following attribution information:

  • symbol.last_editor: The last person who edited the symbol
  • symbol.editor_history: List of all editors who have touched the symbol
  • symbol.is_ai_authored: Boolean indicating if the symbol was authored by AI

Learn More

Contributing

Feel free to submit issues and enhancement requests!